Home Ambitoric geometry I: Einstein metrics and extremal ambikähler structures
Article
Licensed
Unlicensed Requires Authentication

Ambitoric geometry I: Einstein metrics and extremal ambikähler structures

  • Vestislav Apostolov EMAIL logo , David M. J. Calderbank and Paul Gauduchon
Published/Copyright: August 19, 2014

Abstract

We present a local classification of conformally equivalent but oppositely oriented 4-dimensional Kähler metrics which are toric with respect to a common 2-torus action. In the generic case, these “ambitoric” structures have an intriguing local geometry depending on a quadratic polynomial q and arbitrary functions A and B of one variable.

We use this description to classify 4-dimensional Einstein metrics which are hermitian with respect to both orientations, as well as a class of solutions to the Einstein–Maxwell equations including riemannian analogues of the Plebański–Demiański metrics. Our classification can be viewed as a riemannian analogue of a result in relativity due to R. Debever, N. Kamran, and R. McLenaghan, and is a natural extension of the classification of selfdual Einstein hermitian 4-manifolds, obtained independently by R. Bryant and the first and third authors.

These Einstein metrics are precisely the ambitoric structures with vanishing Bach tensor, and thus have the property that the associated toric Kähler metrics are extremal (in the sense of E. Calabi). Our main results also classify the latter, providing new examples of explicit extremal Kähler metrics. For both the Einstein–Maxwell and the extremal ambitoric structures, A and B are quartic polynomials, but with different conditions on the coefficients. In the sequel to this paper we consider global examples, and use them to resolve the existence problem for extremal Kähler metrics on toric 4-orbifolds with second Betti number b2=2.

Award Identifier / Grant number: Discovery Grant

Funding statement: The first author was supported by an NSERC Discovery Grant and is grateful to the Institute of Mathematics and Informatics of the Bulgarian Academy of Sciences where a part of this project was realized. The second author thanks the Leverhulme Trust and the William Gordon Seggie Brown Trust for a fellowship when this project was conceived in 2001, and to the EPSRC for a subsequent Advanced Research Fellowship.

A The projective line and transvectants

Let 𝕎 be a 2-dimensional real vector space equipped with a symplectic form κ (a nonzero element of 2𝕎*). This defines an isomorphism 𝕎𝕎* sending u𝕎 to the linear form u:vκ(u,v); similarly there is a Lie algebra isomorphism from 𝔰𝔩(𝕎) (the trace-free endomorphisms of 𝕎) to S2𝕎* (the quadratic forms on 𝕎, under Poisson bracket {,}) sending a𝔰𝔩(𝕎) to the quadratic form uκ(a(u),u).

The quadratic form -det on 𝔰𝔩(𝕎) induces a quadratic form Q on S2𝕎* proportional to the discriminant, which polarizes to give an 𝔰𝔩(𝕎)-invariant inner product

p,p~=Q(p+p~)-Q(p)-Q(p~)

of signature (2,1) satisfying the following identity:

(A.1)Q({p,p~})=p,p~2-4Q(p)Q(p~).

The analysis can be made more explicit by introducing a symplectic basis e1,e2 of 𝕎 (so that κ(e1,e2)=1) and hence an affine coordinate z on 𝐏(𝕎) (with [w]=[z([w])e1+e2]). A quadratic form qS2𝕎* may then be written

q(z)=q0z2+2q1z+q2

with polarization

q(x,y)=q0xy+q1(x+y)+q2.

In these coordinates the Poisson bracket of q(z) with w(z) is

{q,w}(z)=q(z)w(z)-w(z)q(z)

with

{q,w}0=2q0w1-2q1w0,{q,w}1=q0w2-q2w0,{q,w}2=2q1w2-2q2w1,

and the quadratic form and inner product on S2𝕎* are

Q(q)=q12-q0q2andq,p=2q1p1-(q2p0+q0p2).

Elements of the mth symmetric tensor power Sm𝕎* may similarly be viewed as polynomials in one variable of degree at most m. The tensor product Sm𝕎*Sn𝕎*, for n,m, has the following Clebsch–Gordan decomposition into irreducibles:

(A.2)Sm𝕎*Sn𝕎*=r=0min{m,n}Sm+n-2r𝕎*.

For any r=0,,min{m,n}, the corresponding SL(𝕎)-equivariant map

Sm𝕎*Sn𝕎*Sm+n-2r𝕎*

(well-defined up to a multiplicative constant) is called the transvectant of order r, and denoted by (p,q)(r)—see e.g. Olver [39]. For m=n, the transvectant of order r is symmetric if r is even, and skew if r is odd. When p,q are regarded as polynomials in one variable, it may be written explicitly as

(A.3)(p,q)(r)=j=0r(-1)j(n-jr-j)(m-r+jj)p(j)q(r-j),

where p(j) stands for the j-th derivative of p, with p(0)=p, and similarly for q(r-j). In particular, (p,q)(0) is multiplication, and for any p,qS2𝕎*, (p,q)(1) and (p,q)(2) are constant multiples of the Poisson bracket and inner product respectively.

Elements of Sm𝕎* (and corresponding polynomials in an affine coordinate) may be viewed as (algebraic) sections of the degree m line bundle 𝒪(m) over 𝐏(𝕎); in particular, there is a tautological section of 𝒪(1)𝕎. Formula (A.3) for transvectants extends from algebraic sections to general smooth sections.

B Killing tensors and ambitoric conformal metrics

The material in this appendix is related to work of W. Jelonek [30, 29, 31] and some well-known results in general relativity, see [19] and [32]. To provide a different slant, we take a conformal viewpoint (cf. [15, 17, 22, 45]) and make explicit the connection with M. Pontecorvo’s description [43] of hermitian structures which are conformally Kähler. We then specialize the analysis to ambitoric structures.

B.1 Conformal Killing objects

Let (M,c) be a conformal manifold. Among the conformally invariant linear differential operators on M, there is a family which are overdetermined of finite type, sometimes known as twistor or Penrose operators; their kernels are variously called twistors, tractors, or other names in special cases. Among the examples where the operator is first order are the equations for twistor forms (also known as conformal Killing forms) and conformal Killing tensors, both of which include conformal vector fields as a special case. There is also a second order equation for Einstein metrics in the conformal class. Apart from the obvious presence of (conformal) Killing vector fields and Einstein metrics, conformal Killing 2-tensors and twistor 2-forms are very relevant to the present work.

Let S0kTM denote the bundle of symmetric (0,k)-tensors 𝒮0 which are tracefree with respect to c in the sense that

i𝒮0(εi,εi,)=0

for any conformal coframe εi. In particular, for k=2, 𝒮0S02TM may be identified with σ0L2Sym0(TM) via

ασ0(X)=𝒮0(α,c(X,))

for any 1-form α and vector field X. Here Sym0(TM) is the bundle of tracefree endomorphisms of TM which are symmetric with respect to c; thus σ0 satisfies

c(σ0(X),Y)=c(X,σ0(Y))

and hence defines a (weighted) (2,0)-tensor S0 in L4S02T*M, another isomorph of S02TM (in the presence of c).

A conformal Killing (2-)tensor is a section 𝒮0 of S02TM such that the section sym0D𝒮0 of L-2S03TM is identically zero, where D is any Weyl connection (such as the Levi-Civita connection of any compatible metric) and sym0 denotes orthogonal projection onto L-2S03TM inside T*MS2TML-2TMS2TM. Equivalently

symD𝒮0=sym(χc)

for some vector field χ. Taking a trace, we find that

(n+2)χ=2δD𝒮0,

where δD𝒮0 denotes trcD𝒮0, which may be computed, using a conformal frame ei with dual coframe εi, as iDei𝒮0(εi,). Thus 𝒮0 is conformal Killing if and only if

(B.1)symD𝒮0=2n+2sym(cδD𝒮0).

This is independent of the choice of Weyl connection D. On the open set where 𝒮0 is nondegenerate, there is a unique such D with δD𝒮0=0, and hence a nondegenerate 𝒮0 is conformal Killing if and only if there is a Weyl connection D with symD𝒮0=0.

A conformal Killing 2-form is a section ϕ of L32T*M such that π(Dϕ)=0 (for any Weyl connection D) where π is the projection orthogonal to L33T*M and LT*M in T*ML32T*M. It is often more convenient to identify ϕ with a section Φ of L𝔰𝔬(TM) via ϕ(X,Y)=c(Φ(X),Y), where 𝔰𝔬(TM) denotes the bundle of skew-symmetric endomorphisms of TM with respect to c.

B.2 Conformal Killing tensors and complex structures

In four dimensions a conformal Killing 2-form splits into selfdual and antiselfdual parts Φ±, which are sections of L𝔰𝔬±(TM)L3±2T*M. Following M. Pontecorvo [43], nonvanishing conformal Killing 2-forms Φ+ and Φ- describe oppositely oriented Kähler metrics in the conformal class, by writing Φ±=±J±, where ± are sections of L and J± are oppositely oriented complex structures: the Kähler metrics are then g±=±-2c. Conversely if (g±=±-2c,J±) are Kähler and D± denote the Levi-Civita connections of g±, then D±(±J±)=0 so Φ±=±J± are conformal Killing 2-forms.

The tensor product of sections Φ+ and Φ- of L𝔰𝔬+(TM) and L𝔰𝔬-(TM) defines a tensor Φ+Φ-: as a section of L2Sym0(TM), this is the composite (Φ+Φ-=Φ-Φ+); as a section of L4S02T*M it satisfies (Φ+Φ-)(X,Y)=c(Φ+(X),Φ-(Y)).

When Φ±=±J± are nonvanishing, then Φ+Φ-=+-J+J- is a symmetric endomorphism with two rank 2 eigenspaces at each point. Conversely if σ0 is such a symmetric endomorphism, we may write σ0=2J+J- for uniquely determined almost complex structures J± up to overall sign, and a positive section of L.

Proposition 14

A nonvanishing section σ0=2J+J- of L2Sym0(TM) (as above) is associated to a conformal Killing 2-tensor S0 if and only if J± are integrable complex structures which are “Kähler on average”[4] with length scale , in the sense that if D± denote the canonical Weyl connections of J±, then the connection D=12(D++D-) preserves the length scale (i.e., D++D-=0).

If these equivalent conditions hold, then also symDS0=0.

Proof.

Let D, D+, D- be Weyl connections with

D=12(D++D-)

in the affine space of Weyl connections. (Thus the induced connections on L are related by D=D++θ=D--θ for some 1-form θ.) Straightforward calculation shows that

Dσ0=D(2)J+J-+2(D+J+J-+J+D-J-)+R,

where R is an expression (involving θ) whose symmetrization vanishes (once converted into a trilinear form using c). If J± are integrable and Kähler on average, then taking D± to be the canonical Weyl connections and the preferred length scale, 2J+J- is thus associated to a conformal Killing tensor 𝒮0 with symD𝒮0=0.

For the converse, it is convenient (for familiarity of computation) to work with the associated (2,0)-tensor S0 with S0(X,Y)=2c(J+J-X,Y). Since S0 is nondegenerate, and associated to a conformal Killing tensor, we can let D=D+=D- be the unique Weyl connection with symDS0=0: note that sym:L4T*MS2T*ML4S3T*M here becomes the natural symmetrization map. Thus

X,Y,ZDX(2)c(J+J-Y,Z)=X,Y,Z2(c((DXJ+)J-Y,Z)+c(J+(DXJ-)Y,Z)),

where the sum is over cyclic permutations of the arguments. If X,Y,Z belong to a common eigenspace of S0, then the right hand side is zero—this follows because, for instance, c((DXJ±)J±Y,Z) is skew in Y and Z whereas the cyclic sum of the two terms is totally symmetric.

It follows that D=0, hence the right hand side is identically zero in X,Y,Z. Additionally c(DXJ±,) is J±-anti-invariant. Thus these 2-forms vanish when their arguments have opposite types ((1,0) and (0,1)) with respect to the corresponding complex structure. Now suppose for example that Z1 and Z2 have type (1,0) with respect to J+, but opposite types with respect to J- (J+ and J- are simultaneously diagonalizable on TM). Then by substituting first X=Y=Z1, Z=Z2 into

X,Y,Zc((DXJ+)J-Y,Z)=X,Y,Zc((DXJ-)Y,J+Z),

and then X=Y=Z2, Z=Z1, we readily obtain

c((DZ1J+)Z1,Z2)=0=c((DZ2J+)Z1,Z2).

Thus DJ+XJ+=J+DXJ+ for all X and J+ is integrable. Similarly, we conclude J- is integrable. ∎

Since D is the Levi-Civita connection Dg of the “barycentric” metric g=-2c, it follows that S0=g(J+J-,) is a Killing tensor with respect to g, i.e., satisfies symDgS0=0 if and only if J+ and J- are integrable and Kähler on average, with barycentric metric g. More generally, we can use this result to characterize, for any metric g in the conformal class and any functions f,h, the case that

(B.2)S(,)=fg(,)+hg(J+J-,)

is a Killing tensor with respect to g. If θ± are the Lee forms of (g,J±), i.e., D±=Dg±θ±, then we obtain the following more general corollary.

Corollary 3

The tensor S=fg+hg(J+J-,), with h nonvanishing, is a Killing tensor with respect to g if and only if

(B.3)J+ and J- are both integrable,
(B.4)θ++θ-=-dhh,
(B.5)J+df=J-dh.

(Obviously when h is identically zero, S is a Killing tensor if and only if f is constant.)

B.3 Conformal Killing tensors and the Ricci tensor

Let

ric0g=ricg-1nsgg

be the tracefree part of the Ricci tensor of a compatible metric g=μg-2c on a conformal n-manifold (M,c). Then, the section 𝒮0g of S02TM, corresponding to the section μg4ric0g of L4S02T*M, is

𝒮0g(α,β)=ric0g(α,β)

(where for αT*M, g(α,)=α)).

The differential Bianchi identity implies that

0=δg(ricg-12sgg)=δgric0g-n-22ndsg.

Hence the following are equivalent:

  1. 𝒮0g is a conformal Killing tensor,

  2. one has

    symDg𝒮0g=n-2n(n+2)sym(g-1dsg),
  3. ricg-2n+2sgg is a Killing tensor with respect to g,

  4. for all vector fields X,

    DXgricg(X,X)=2n+2dsg(X)g(X,X).

Riemannian manifolds (M,g) satisfying these conditions were introduced by A. Gray as 𝒜C-manifolds [24]. Relevant for this paper is the case n=4 and the assumption that ricg has two rank 2 eigendistributions, which has extensively been studied by W. Jelonek [29, 31].

Supposing that g is not Einstein, Corollary 3 implies, as shown by Jelonek, that

ricg-13sgg=fg+hg(J+J-,)

is Killing with respect to g if and only if (B.3)–(B.5) are satisfied. Since J± are both integrable, Jelonek refers to such manifolds as bihermitian Gray surfaces. It follows from [7] that both (g,J+) and (g,J-) are conformally Kähler, so that in the context of the present paper, a better terminology would be ambikähler Gray surfaces.

However, the key feature of such metrics is that the Ricci tensor is J±-invariant: as long as J± are conformally Kähler, Proposition 11 applies to show that the manifold is either ambitoric or of Calabi type; it is not necessary that the J±-invariant Killing tensor constructed in the proof is equal to the Ricci tensor ricg.

Jelonek focuses on the case that the ambihermitian structure has Calabi type. This is justified by the global arguments he employs. In the ambitoric case, there are strong constraints, even locally.

B.4 Killing tensors and hamiltonian 2-forms

The notion of hamiltonian 2-forms on a Kähler manifold (M,g,J,ω) has been introduced and extensively studied in [2, 3]. According to [3], a J-invariant 2-form ϕ is hamiltonian if it satisfies

(B.6)DXϕ=12(dσJX-JdσX)

for any vector field X, where X=g(X) and σ=trωϕ=g(ϕ,ω) is the trace of ϕ with respect to ω. An essentially equivalent (but not precisely the same) definition was given in the 4-dimensional case in [2], by requiring that a J-invariant 2-form φ is closed and its primitive part φ0 satisfies

(B.7)DXφ0=-12dσ(X)ω+12(dσJX-JdσX)

for some smooth function σ. In order to be closed, φ must have the form 32σω+φ0.

The relation between the two definitions is straightforward: φ=32σω+φ0 is closed and verifies (B.7) if and only if ϕ=φ0+12σω satisfies (B.6).

Specializing Corollary 3 to the case when the metric g is Kähler with respect to J=J+ allows us to identify J-invariant symmetric Killing tensors with hamiltonian 2-forms as follows:

Proposition 15

Let (M,g,J,ω) be a Kähler surface, let S be a symmetric J-invariant tensor on (M,g,J,ω), and let ψ(,)=S(J,) be the associated J-invariant 2-form. Then S is Killing if and only if ϕ=ψ-(trωψ)ω is a hamiltonian 2-form (i.e., verifies (B.6)).

Proof.

As observed in [3, p. 407], ϕ satisfies (B.6) if and only if φ=ϕ+(trωϕ)ω is a closed 2-form and ψ=ϕ-(trωϕ)ω is the 2-form associated to a J-invariant Killing tensor (this is true in any complex dimension m>1).

Noting that the 2-forms φ and ψ are related by

φ=ψ-2trωψm-1ω,

we claim that in complex dimension m=2, the 2-form φ=ϕ-2(trωψ)ω is automatically closed, provided that ψ is the 2-form associated to a J-invariant Killing tensor S. Indeed, under the Kähler assumption the conditions (B.3)–(B.4) specialize as

(B.8)J- is integrable,
(B.9)θ-=-dhh,

It follows that (g-=h-2g,J-,ω-=g-(J-,)) is Kähler. From (B.2) we have

(B.10)ψ=fω++h3ω-,

where ω+=g(J+,) denotes the Kähler form of (g,J+). In particular, the trace of φ with respect to ω+ is equal to 2f while the condition (B.5) and the fact that ω- is closed imply that φ=ψ-4fω+=-3fω++h3ω- is closed too. ∎

B.5 Killing tensors associated to ambitoric structures

We have seen in the previous subsections that there is a link between Killing tensors and ambihermitian structures. We now make this link more explicit in the case of ambitoric metrics.

In the ambitoric situation, the barycentric metric g0 (see Section 4) satisfies

θ+0+θ-0=0.

It then follows from Corollary 3 that the (tracefree) symmetric bilinear form g0(I,) (with I=J+J-) is Killing with respect to g0. More generally, let g be any (K1,K2)-invariant riemannian metric conformal to g0, so that g can be written as g=hg0 for some positive function h(x,y), where x,y are the coordinates introduced in Section 4. Then

θ+g+θ-g=-dlogh.

From Corollary 3 again, the symmetric bilinear form S0(,)=hg(I,) is conformal Killing. Moreover, by condition (B.5) in Proposition 3, it can be completed into a Killing symmetric bilinear form S=fg+S0 if and only if the 1-form dhI is closed. Since Idx=-dx and Idy=dy, it follows that dhI is closed if and only if hxdx-hydy is closed, if and only if hxy=0; the general solution is h(x,y)=F(x)-G(y), for some functions F,G. Note that the coefficient f(x,y) is determined by df=-Idh=F(x)dx+G(y)dy (see (B.5)), so we can take without loss f(x,y)=F(x)+G(y). Thus, S is Killing, with eigenvalues 2F(x) and 2G(y).

A similar argument shows that any metric of the form g=f(z)g0, where g0 is the barycentric metric of an ambikähler pair of Calabi type and z is the momentum coordinate introduced in Section 3.2, admits a nontrivial symmetric Killing tensor of the form

S(,)=f(z)g(,)+f(z)g(I,)

(and hence with eigenvalues (2f(z),0)).

It follows that there are infinitely many 𝔱-invariant metrics in a given ambitoric conformal class, which admit nontrivial symmetric Killing tensors.

There are considerably fewer such metrics with ambihermitian Ricci tensor. By Proposition 13, these have the form g=hg0, where h(x,y)=(x-y)q(x,y)/p(x,y)2. In order for g to admit a nontrivial symmetric Killing tensor, we must have hxy=0. A calculation shows that this happens if and only if Q(p)=0 (i.e., p(z) has repeated roots). Since p is orthogonal to q, this can only happen if Q(q)0 and there are generically (Q(q)>0) just two solutions for p, which coincide if Q(q)=0.

Acknowledgements

The authors are grateful to Liana David and the Centro Georgi, Pisa, and to Banff International Research Station for opportunities to meet in 2006 and 2009, when much of this work was carried out. They thank Maciej Dunajski, Niky Kamran, Claude LeBrun and Arman Taghavi-Chabert for very useful discussions and comments.

References

[1] Abreu M., Kähler geometry of toric varieties and extremal metrics, Internat. J. Math. 9 (1998), 641–651. 10.1142/S0129167X98000282Search in Google Scholar

[2] Apostolov V., Calderbank D. M. J. and Gauduchon P., Weakly self-dual Kähler surfaces, Compos. Math. 135 (2003), 279–322. 10.1023/A:1022251819334Search in Google Scholar

[3] Apostolov V., Calderbank D. M. J. and Gauduchon P., Hamiltonian 2-forms in Kähler geometry I: General theory, J. Differential Geom. 73 (2006), 359–412. 10.4310/jdg/1146169934Search in Google Scholar

[4] Apostolov V., Calderbank D. M. J. and Gauduchon P., Ambitoric geometry II: Extremal toric surfaces and Einstein 4-orbifolds, preprint 2013, http://arxiv.org/abs/1302.6979. 10.24033/asens.2266Search in Google Scholar

[5] Apostolov V., Calderbank D. M. J., Gauduchon P. and Tønnesen-Friedman C., Hamiltonian 2-forms in Kähler geometry II: Global classification, J. Differential Geom. 68 (2004), 277–345. 10.4310/jdg/1115669513Search in Google Scholar

[6] Gauduchon V. Apostolov, D. M. J. Calderbank, P. and Tønnesen-Friedman C., Extremal Kähler metrics on ruled manifolds and stability, Géométrie différentielle, physique mathématique, mathématiques et socété (II), Astérisque 322, Société Mathématique de France, Paris (2008), 93–150. Search in Google Scholar

[7] Apostolov V. and Gauduchon P., The Riemannian Goldberg–Sachs theorem, Internat. J. Math. 8 (1997), 421–439. 10.1142/S0129167X97000214Search in Google Scholar

[8] Apostolov V. and Gauduchon P., Self-dual Einstein Hermitian 4-manifolds, Ann. Sc. Norm. Super. Pisa Cl. Sci. (5) 1 (2002), 203–243. Search in Google Scholar

[9] Bérard-Bergery L., Sur de nouvelles variétés riemanniennes d’Einstein, Inst. Élie Cartan 6 (1982), 1–60. Search in Google Scholar

[10] Besse A. L., Einstein manifolds, Ergeb. Math. Grenzgeb. (3) 10, Springer-Verlag, Berlin 1987. 10.1007/978-3-540-74311-8Search in Google Scholar

[11] Boyer C. P., Conformal duality and compact complex surfaces, Math. Ann. 274 (1986), 517–526. 10.1007/BF01457232Search in Google Scholar

[12] Bryant R., Bochner–Kähler metrics, J. Amer. Math. Soc. 14 (2001), 623–715. 10.1090/S0894-0347-01-00366-6Search in Google Scholar

[13] Calabi E., Extremal Kähler metrics, Seminar on differential geometry, Ann. of Math. Stud. 102, Princeton University Press, Princeton (1982), 259–290. 10.1515/9781400881918-016Search in Google Scholar

[14] Calabi E., Extremal Kähler metrics II, Differential geometry and complex analysis, Springer-Verlag, Berlin (1985), 95–114. 10.1007/978-3-642-69828-6_8Search in Google Scholar

[15] Calderbank D. M. J. and Diemer T., Differential invariants and curved Bernstein–Gel’fand–Gel’fand sequences, J. reine angew. Math. 537 (2001), 67–103. 10.1515/crll.2001.059Search in Google Scholar

[16] Calderbank D. M. J. and Pedersen H., Selfdual spaces with complex structures, Einstein–Weyl geometry and geodesics, Ann. Inst. Fourier (Grenoble) 50 (2000), 909–951. 10.5802/aif.1779Search in Google Scholar

[17] Čap A., Slovák J. and Souček V., Bernstein–Gelfand–Gelfand sequences, Ann. of Math. (2) 154 (2001), no. 1, 97–113. 10.2307/3062111Search in Google Scholar

[18] Chave T., Pedersen H., Tønnesen-Friedman C. and Vallent G., Extremal Kähler metrics and Hamiltonian functions, J. Geom. Phys. 31 (1999), 25–34. 10.1016/S0393-0440(98)00069-2Search in Google Scholar

[19] Cosgrove C., A new formulation of the field equations for the stationary axisymmetric gravitational field II. Separable solutions, J. Phys. A 11 (1928), 2405–2430. 10.1088/0305-4470/11/12/008Search in Google Scholar

[20] Debever R., Kamran N. and McLenaghan R. G., Exhaustive integration and a single expression for the general solution of the type D vacuum and electrovac field equations with cosmological constant for a non-singular aligned Maxwell field, J. Math. Phys. 25 (1984), 1955–1972. 10.1063/1.526386Search in Google Scholar

[21] Derdziński A., Self-dual Kähler manifolds and Einstein manifolds of dimension four, Compos. Math. 49 (1983), 405–433. Search in Google Scholar

[22] Gauduchon P., Structures de Weyl et théorèmes d’annulation sur une variété conforme autoduale, Ann. Sc. Norm. Super. Pisa Cl. Sci. (4) 18 (1991), 563–629. Search in Google Scholar

[23] Goldberg J. N. and Sachs R. K., A theorem on Petrov types, Acta Phys. Polon. 22 (1962), 13–23. Search in Google Scholar

[24] Gray A., Einstein-like manifolds which are not Einstein, Geom. Dedicata 7 (1978), 259–280. 10.1007/BF00151525Search in Google Scholar

[25] Griffiths G. B. and Podolsky J., A new look at the Plebański–Demiański family of solutions, Internat. J. Modern Phys. D 15 (2006), 335–370. 10.1142/S0218271806007742Search in Google Scholar

[26] Guillemin V., Kähler structures on toric varieties, J. Differential Geom. 40 (1994), 285–309. 10.4310/jdg/1214455538Search in Google Scholar

[27] Hwang A. and Simanca S., Distinguished Kähler metrics on Hirzebruch surfaces, Trans. Amer. Math. Soc. 347 (1995), 1013–1021. 10.1090/S0002-9947-1995-1246528-9Search in Google Scholar

[28] Hwang A. D. and Singer M. A., A momentum construction for circle-invariant Kähler metrics, Trans. Amer. Math. Soc. 354 (2002), 2285–2325. 10.1090/S0002-9947-02-02965-3Search in Google Scholar

[29] Jelonek W., Bi-Hermitian Gray surfaces, Pacific J. Math. 222 (2005), 57–68. 10.1016/j.difgeo.2008.06.006Search in Google Scholar

[30] Jelonek W., Compact Kähler surfaces with harmonic anti-self-dual Weyl tensor, Differential Geom. Appl. 16 (2006), 267–276. 10.1016/S0926-2245(02)00076-1Search in Google Scholar

[31] Jelonek W., Bi-Hermitian Gray surfaces. II, Differential Geom. Appl. 27 (2009), 64–74. 10.1016/j.difgeo.2008.06.006Search in Google Scholar

[32] Kamran N. and McLenaghan R. G., Separation of variables and constants of the motion for the Dirac equation on curved spacetime, Acad. Roy. Belg. Bull. Cl. Sci. (5) 70 (1984), no. 10, 596–610. 10.3406/barb.1984.72378Search in Google Scholar

[33] Kühnel W., Conformal transformations between Einstein spaces, Conformal geometry (Bonn 1985/1986), Aspects Math. 12, Vieweg-Verlag, Braunschweig (1988), 105–146. 10.1007/978-3-322-90616-8_5Search in Google Scholar

[34] LeBrun C. R., Explicit self-dual metrics on P2##P2, J. Differential Geom. 34 (1991), 223–253. 10.4310/jdg/1214446999Search in Google Scholar

[35] LeBrun C. R., Einstein metrics on complex surfaces, Geometry and physics (Aarhus 1995), Lect. Notes Pure Appl. Math. 184, Dekker, New York (1997), 167–176. 10.1201/9781003072393-11Search in Google Scholar

[36] LeBrun C. R., Einstein–Maxwell equations, extremal Kähler metrics, and Seiberg–Witten theory, The many facets of geometry. A tribute to Nigel Hitchin, Oxford University Press, Oxford (2009), 17–33. 10.1093/acprof:oso/9780199534920.003.0003Search in Google Scholar

[37] Legendre E., Toric geometry of convex quadrilaterals, J. Symplectic Geom. 9 (2011), 343–385. 10.4310/JSG.2011.v9.n3.a3Search in Google Scholar

[38] Nurowski P., Einstein equations and Cauchy–Riemann geometry, PhD thesis, SISSA, 1993, http://digitallibrary.sissa.it/handle/1963/31. Search in Google Scholar

[39] Olver P. J., Classical invariant theory, London Math. Soc. Stud. Texts 44, Cambridge University Press, Cambridge 1999. 10.1017/CBO9780511623660Search in Google Scholar

[40] Page D., A compact rotating gravitational instanton, Phys. Lett. B 79 (1978), 235–238. 10.1142/9789814539395_0035Search in Google Scholar

[41] Penrose R. and Rindler W., Spinors and spacetime. Vol. 2, Cambridge University Press, Cambridge 1986. 10.1017/CBO9780511524486Search in Google Scholar

[42] Plebański J. F. and Demiański M., Rotating, charged, and uniformly accelerating mass in general relativity, Ann. Phys. 98 (1976), 98–127. 10.1016/0003-4916(76)90240-2Search in Google Scholar

[43] Pontecorvo M., On twistor spaces of anti-self-dual Hermitian surfaces, Trans. Amer. Math. Soc. 331 (1992), 653–661. 10.1090/S0002-9947-1992-1050087-0Search in Google Scholar

[44] Przanowski M. and Broda B., Locally Kähler gravitational instantons, Acta Phys. Polon. B 14 (1983), 637–661. Search in Google Scholar

[45] Semmelmann U., Conformal Killing forms on Riemannian manifolds, Math. Z. 245 (2003), 503–527. 10.1007/s00209-003-0549-4Search in Google Scholar

[46] Tønnesen-Friedman C., Extremal Kähler metrics on minimal ruled surfaces, J. reine angew. Math. 502 (1998), 175–197. 10.1515/crll.1998.086Search in Google Scholar

[47] Tønnesen-Friedman C., Extremal Kähler metrics and Hamiltonian functions. II, Glasg. Math. J. 44 (2002), 241–253. 10.1017/S0017089502020050Search in Google Scholar

Received: 2013-3-4
Revised: 2013-10-28
Published Online: 2014-8-19
Published in Print: 2016-12-1

© 2016 by De Gruyter

Downloaded on 24.9.2025 from https://www.degruyterbrill.com/document/doi/10.1515/crelle-2014-0060/html
Scroll to top button